On Fri, 2008-10-24 at 13:45 +0100, sebb wrote:
> On 24/10/2008, Lovette, Steve <[EMAIL PROTECTED]> wrote:
> > Sebbaz
> >   From what I have read the use of algorithms that have been shown to be
> >  breakable become unacceptable. There is literature on the web about
> >  this. From reading the government NIST web site and the government STIGs
> >  that recommend only the SHA-x algorithms to be used in sensitive
> >  applications. MD5 is not a government approved algorithm to be used in
> >  hashing functions where encryption is involved.
> 
> OK, but so what?
> 
> >  That said your point about HTTP client may well be the best counter
> >  point. Since HTTP client runs on the client and the client is always
> >  suspect then perhaps this is a sufficient argument.
> 
> I think you still misunderstand what HC is for.
> 
> It is irrelevant where HC runs; the point is that it is a client
> library, i.e. it talks to servers.
> 
> If the server needs MD5 for something, then HC will use that.
> HC does not use MD5 for its own purposes.
> 

Sebastian, et al

I took a brief look at the DIGEST authentication scheme implementation
in HttpClient 4.0 and it appears HttpClient will reject a challenge
unless the specified digest algorithm is either MD5 or MD5-sess. As far
as I can tell these are two algorithms mentioned in RFC 2617. There is
no mentioning of SHA-x algorithms in the spec. However, it would
certainly make sense to ensure HttpClient can support alternative
digests, if the server requests an algorithm other than MD5 or MD5-sess.

Steve, 

If you think the present implementation of DIGEST authentication is not
secure enough, feel free to open a JIRA for this issue

https://issues.apache.org/jira/browse/HTTPCLIENT 

Oleg



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to